Do Brazilians Speak English?

The 5% is not what it is credited as

Trace it and the attribution shifts. The number appears in a report published by the British Council, but the research behind it was conducted by a private market-research institute the British Council commissioned. So "the British Council found that 5% of Brazilians speak English" moves the finding one step from where it happened.

What was measured matters more. The respondents declared having some knowledge of English — self-assessment, not a test. And the sample was not the Brazilian population: it was 720 people, aged 18 to 55, from the middle and lower-middle class, who were already taking an English course or intending to.

That last filter is the one that breaks the statistic. A survey of people enrolling in language courses cannot produce a rate for the country, in either direction.

⚠️ A second claim travels with it: "only 1% are fluent." We searched the full report — with positive and negative controls, because a broken text extraction returns zero for everything and looks like a finding. The words for fluent, advanced and intermediate do not appear. The 1% figure is not in that document.

That is a statement about this report, which is where the claim is usually sourced. It is not a claim that nobody has ever said it.

The other number is published by a language school

The English Proficiency Index is the second source everyone cites. It is produced by a company that sells English courses, from results of its own free online test.

The company is candid about the limitation, in its own words: the test-taking population is self-selected and not guaranteed to be representative. People who take a free online English test are people interested in English — which biases the result upward, not down.

Read with that caveat, its 2025 edition places Brazil 75th of 123 countries and regions, in the band it labels low proficiency, on a score that is a three-year rolling average rather than a single year's measurement.

Nobody official measures it

Here is the gap that explains why the weak numbers survive: Brazil does not collect this.

The census and the national household survey do ask about language — but about indigenous and quilombola languages, which is a different question with a different purpose. Foreign-language ability is not among the topics.

So there is no state measurement to displace the private ones. A 2013 commissioned survey and a language school's index fill a vacuum, and get quoted as though they were census data.

What to plan for

Given all that, the honest advice is behavioural rather than numerical.

Plan as though English will not be available, and treat it as a bonus where it is. Airports, upmarket hotels and parts of the tourist trade are the reliable exceptions. Public services, transport counters, pharmacies and most restaurants are not — and none of the sources above suggests otherwise, whatever their flaws.

The practical consequence is that the things which normally get sorted by asking someone need sorting another way. That is the argument for reading how the payment system works before you need it, for knowing what the cashier is asking when they want your tax number, and for understanding what a Brazilian ATM screen means rather than guessing at a counter. Same for buying a SIM card, reading an address and reading a phone number, which are formats rather than conversations. In an emergency it matters more: what the safety numbers actually say includes the fact that English-language response is not guaranteed. And whether you need a visa is settled in writing, before anyone has to speak.

Sources

• The commissioned report published by the British Council — the 5.1% figure, what it measured, who researched it, and the sample composition. Accessed September 3, 2026.
• EF English Proficiency Index 2025, full report — Brazil's position and band, and the skill-level scores. Accessed September 3, 2026.
• EF, about the index — the publisher's own statement that the test-taking population is self-selected and not guaranteed to be representative, and the sample size. Accessed September 3, 2026.
• IBGE, Continuous National Household Sample Survey and 2022 Census — checked for foreign-language ability; neither asks it. Accessed September 3, 2026.

What we could not verify

• Any representative measurement of English ability in Brazil. We found none, from any source, official or private.
• Whether the "1% fluent" claim originates in some other document. We established it is not in the report it is usually attributed to; we did not trace every possible origin.
• Regional or age breakdowns. Both available sources have samples too narrow to support them, and neither publishes such a cut.